Quickstart Guide: zzpower

Overview

zzpower is an interactive Shiny application for power analysis and sample size calculations in clinical trials. The package supports multiple statistical tests through a scalable plugin architecture.

Installation

# Install from GitHub
devtools::install_github("rgt47/zzpower")

library(zzpower)

Basic Usage

Launch the interactive application:

launch_zzpower()

The application opens in your web browser with a tabbed interface for different statistical tests.

Supported Statistical Tests

Two-Group t-test (Independent Samples)

For parallel-group clinical trial designs:

Effect size methods: Cohen’s d, percent reduction, difference in scores
Design features: dropout rates, unequal allocation ratios
Output: power curves, sample size tables, downloadable reports

Paired t-test

For before-after or crossover designs:

Effect size: Standardized mean difference
Sample size calculation for matched pairs

One-Sample t-test

For comparison against a fixed reference value:

Effect size: Cohen’s d
Single-group sample size determination

Two Proportions

For binomial outcome comparisons:

Effect size methods: two proportions, difference, odds ratio, relative risk
Sample size calculation for binary endpoints

Correlation Test

For bivariate relationship testing:

Effect size: correlation coefficient (r)
Sample size for detecting correlations

Workflow Example

1. Launch Application

launch_zzpower()

2. Select Test Type

Choose the appropriate tab for your study design (e.g., “Two-Group t-test”).

3. Set Sample Size

Use the slider to set your planned total sample size.

4. Specify Effect Size

Choose an effect size method and set the range to evaluate:

Cohen’s d: Direct standardized effect (0.2 = small, 0.5 = medium, 0.8 = large)
Percent Reduction: Effect as percentage reduction from control
Difference in Scores: Absolute difference in outcome units

5. Configure Design

Adjust advanced settings:

Dropout/drop-in rates
Allocation ratio (1:1, 2:1, etc.)
Alpha level (0.05, 0.01, etc.)
One-sided vs two-sided testing

6. Review Results

Examine the interactive outputs:

Power curve showing power vs effect size
Results table with exact values
Study design summary

7. Download Report

Generate a report documenting your power analysis in text or HTML format.

Function Reference

Function	Purpose
`launch_zzpower()`	Launch interactive Shiny app

Key Parameters for launch_zzpower()

Parameter	Description	Default
`port`	Port number	NULL (random)
`host`	IP address	“127.0.0.1”
`launch.browser`	Open browser	TRUE

Architecture

zzpower uses a scalable plugin/registry architecture:

Registry: Central specification of all test definitions
Generic UI Builder: Dynamic UI generation from specifications
Generic Server Factory: Dynamic server logic from specifications

This architecture enables support for 20+ statistical tests with minimal code duplication.

Dependencies

shiny: Web application framework
bslib: Bootstrap 5 theming
pwr: Power analysis calculations
ggplot2: Visualization
DT: Interactive tables

Next Steps

For detailed information on the framework architecture, see the package documentation and CLAUDE.md.

Ronald (Ryy) G. Thomas

2026-05-01